Po-02-016 physiologic signals from crt devices are indicative of crt response in smart-msp study

Cardiac resynchronization therapy (CRT) is a well-established treatment for some heart failure patients with reduced ejection fraction, but a subset of these patients does not respond to CRT. Multi-site pacing (MSP) is proposed to improve clinical response to CRT, but it is unclear which CRT subset would benefit from MSP. Physiologic sensor data from CRT devices may provide insights into this question. To determine whether device based physiologic parameters are indicative of CRT response/non-response and thus identify potential candidates for MSP. SMART-MSP study evaluated 6-month CRT response, defined by improved Clinical Composite Score, in patients with CRT defibrillators (CRT-D). Device parameters considered for this analysis included third (S3) and first (S1) heart sounds, thoracic impedance (ITTI), respiration rate, rapid shallow breathing index (RSBI), activity, and the multi-sensor HeartLogic index. From a randomly selected 60% development set of patients (183 CRT responders and 40 non-responders) with at least 150 days of data, the following statistical measures were calculated for each parameter based on the first 60- and 150-day epoch after CRT-D implant: mean (μ), standard deviation (σ) and change from first 30-day mean to last 30-day mean (Δ). To evaluate the effect of each measure on CRT response, odds ratios were calculated and tested for significance via univariable and multivariable logistic regression. The 60-day analysis (Fig 1A) shows that increases in σ60 for S3, S3/S1, RSBI, respiration rate, and HeartLogic index as well as μ60 for RSBI were significantly associated with reduced odds of CRT response in the univariable model. The only signal to survive multivariate elimination was respiration rate σ60. The 150-day analysis (Fig 1B) shows the same indications and that increases in μ150 for RSBI and HeartLogic index were associated with reduced odds while increases in μ150 and Δ150 for activity were associated with increased odds of CRT response. Both respiration rate σ150 and activity Δ150 survived multivariate elimination. Multiple device-based physiologic parameters significantly differed between CRT responders versus non-responders in the SMART-MSP study. 60-day results indicate that there is potential for device-based parameters to identify patients who may benefit from switching to MSP sooner. However, while useful parameters are apparent early on (60-day window), more parameters become significant over a longer follow-up (150 days).